Regular contributor Professor Caleb Smith continues on our theme of discussing the long reach of Graev and related issues. Les
There were five designated orders last week, but only one worth going into much detail on. Of course, it involved Judge Holmes once more considering some implications of Graev. The other orders involved a taxpayer erroneously claiming the EITC with income earned as an inmate (here); and three orders by Judge Gustafson working with pro se taxpayers: two of which are in the nature of assisting the taxpayer (how file a motion to be recognized as next friend here, and clarifying how to enter evidence here) and one granting summary judgment to the IRS (here). Note parenthetically that in the latter order Judge Gustafson goes out of his way to mention that the IRS approved a 6662(a) penalty in compliance with IRC 6751 [erroneously cited as 7651]. IRC 6751, of course, is the issue du jour, and the focus of today’s post.
Rajagopalan & Kumar, et al. v. C.I.R., Dkt. # 21394-11, 21575-11 [here]
Supervisory Approval: Is it Needed for Every “Reason” Behind the Penalty?
For those that need to catch up, the Procedurally Taxing team has provided a wealth of analysis and insight on the Graev/Chai case developments. For “Graev III” fallout readers are encouraged to visit this, this and this post (to name a few). Judge Holmes in particular has been at the forefront of raising (if not quite resolving) unanswered questions that lurk in the aftermath of Graev III. In Judge Holmes’s most recent order, we see two questions bubble to the surface. One of those issues should only provide a temporary headache to the IRS: the procedural hurdle for the IRS to introduce into evidence that they complied with IRC 6751 if the record has been closed. The other issue, however, could well create a lasting migraine for the IRS: whether the IRS form showing supervisor approval also sufficiently shows approval for the rationale of the penalty. That problem isn’t directly answered in the order, but I think it is the most interesting (and most likely to remain a lasting problem) so we will begin our analysis with it.
Imagine the IRS asserts that a taxpayer understated their tax due by $5500 (with that amount being more than 10% of the total tax due). The IRS issues a Notice of Deficiency that throws the book at the taxpayer with an IRC 6662(a) penalty because of this substantial understatement of income tax and because the taxpayer was negligent. In so doing, the IRS is relying on two separate subsections of IRC 6662 as their legal basis for the penalty’s application: subsections (b)(1) and (b)(2).
Imagine further that the IRS did the right thing and had a supervisor sign-off on the penalty prior to issuing the Notice of Deficiency. Does the supervisor need to approve of both rationales (i.e. (b)(1) and (b)(2))? Or is the fact that the penalty, to some degree, got supervisory approval enough on its own? What if the Tax Court finds that this same taxpayer only understated $4500 in tax on their return? Now only negligence could get the IRS to a 6662(a) penalty: do we need to have proof that the supervisor approved of that ground for raising the penalty?
These are questions that Judge Holmes has raised before, in his concurrence of Graev III. Judge Holmes lays out a parade of horribles beginning on page 45 of the opinion, one of which deals with approval of one, but not two, grounds for an IRC 6662(a) penalty (on page 46, point 4). This made me wonder exactly what the supervisory approval form looks like, and if it sets these points apart. With the sincerely appreciated assistance of frequent PT blogger Carl Smith and lead Graev III attorney (also PT contributor) Frank Agostino, I was able to gaze upon this fabled supervisory approval form, which can be found here. And, sure enough, the form does break down 6662 penalties (to a degree). It breaks down IRC 6662 into four categories: (1) Negligence, (2) Substantial Understatement, (3) all other 6662(b) infractions, and (4) 6662(h). The neatly delineated checkboxes certainly make it seem like a supervisor is only “approving” whichever specific penalty rationale they check yes next to.
Looking to the statute at issue provides little guidance on what “amount” of supervisory approval is needed, only that the “initial determination” is personally approved before making the determination. Taking the above accuracy penalty as an example, one could argue that the penalty needing approval is only IRC 6662(a), so that is all that need be approved broadly. The supervisor has agreed that the penalty should apply and the worry of it being used as a bargaining chip is lessoned. The statute isn’t intended to provide a through legal review of all penalty theories, but only to be sure that they aren’t being applied recklessly as “bargaining chips.”
However, one could just as reasonably argue that the nature of the penalty’s application requires some degree of specificity: the penalty is only applied to the amount of the underpayment attributable to that rationale. If our hypothetical taxpayer understated by $5500, but only $1,000 of it is due to negligence, then you would have two potential penalty values: $1,110 (20% of substantial understatement) or $200 (20% for the portion attributable to negligence). Yes, the penalties arise under the same code section (broadly: 6662(a)), but their calculation depend on the rationale (narrowly: 6662(b)(1) or (2)). Since that leads to two different potential penalty amounts, it would seem (in a sense) to be two different penalties. Certainly, one would think two separate approvals were needed if the penalties were IRC 6662(a) or IRC 6662(h), as they apply two different penalty percentages. Why should it be different if they potentially apply against two different amounts of understatement?
Questions I’m sure Judge Holmes looks forward to in future briefs. Though the intent of IRC 6751 is laudable, the language certainly leaves much to be desired.
In the interest of taxpayer rights, however, I think it is important to note that the IRS has created at least some of these problems on their own. From my perhaps biased perspective, accuracy penalties under IRC 6662(a) are most troublesome when applied “automatically” or with little thought against low-income taxpayers that may simply have had difficulty navigating complicated qualifying child rules. In my practice I deal less with the “bargaining chip” and more with the “punitive” aspect of penalties. We have already seen how reflexively the IRS will slap EITC bans without proper approval or documentation here. There may be reason to believe the IRS is just as reflexive with these IRC 6662(a) penalties. Consideration of the relevant IRM is illustrative:
IRM section 20.1.5.1.4 details “Managerial Approval of Penalties.” It lays out the general requirement of IRC 6751 that supervisory approval is required for assessment of a penalty, and then details two important exceptions (one of which I’ll focus on): there is no need for supervisory approval on penalties that are “automatically calculated through electronic means.”
This, by the IRS interpretation, includes IRC 6662(a) penalties for both negligence and substantial understatement if so determined by AUR… so long as no human employee is actually involved in that AUR determination. In other words, we are to trust that no safeguard is needed when the (badly outdated) computers of the IRS determine that there was negligence on the part of the taxpayer. I would note that it appears that this also applies for campus correspondence exams though that is not immediately clear. IRM 20.1.5.1.4(2)(b) implies as much by referring to IRM 20.1.5.1.4(4) (the exception to human approval provision), but that latter provision only mentions the AUR function.
But wait, there’s more. Per that same IRM, if the taxpayer responds to the letter (or notice of deficiency) proposing the penalty then the IRS needs supervisory approval because now it is out of the realm of machines and into the realm of humans. This would seem to imply that taxpayers only have the protection of IRC 6751 if they are noisy. If they aren’t noisy, the IRS hasn’t violated a right of the taxpayer they failed to assert: the right never existed by virtue of failing to assert it. (Apologies for getting metaphysical on that one.)
Bringing it back to the realm of legal/statutory analysis, this still doesn’t seem quite right. Wasn’t the “initial” determination of the penalty already done prior to the taxpayer responding? Or is that irrelevant because at the computer stage it was not an “initial determination of such assessment” (whatever that means)? Judge Holmes, again, has signaled what he believes to be a coming storm on the “initial determination” question. I have no doubt that, given the sloppiness of the statute and the rather poor procedures in place for the IRS, that question is likely to be litigated.
Other Temporary Problems Addressed in Rajagopalan
Though I have devoted the bulk of this post to the issue of “types” of supervisory approval, most of the designated order actually dealt with a different issue. Luckily it is an issue that should be less of a problem moving forward: the IRS scrambling to get evidence of supervisory approval into the court record when the record’s already closed.
As the docket numbers indicate, these consolidated trials have been going on for quite some time: a child born when the Rajagopalan petition was filed would probably be learning their multiplication tables right now. The supervisory approval requirement of IRC 6751 was in effect well before the Rajagopalan trial and record was closed… so how could the IRS possibly have an excuse to reopen the record at this later date?
Obviously, because of the brave new post-Graev III world we now live in. Judge Holmes notes that the IRS had some reason to anticipate the IRC 6751 issue, but doesn’t seem to fault the IRS too much for that failure. Instead, Judge Holmes lays out the requirements to reopen the record: the late evidence must be (1) not merely cumulative or impeaching, (2) material to the issues involved, and (3) likely to change the outcome of the case. In other words, it must be very important towards proving what is at stake (not simply disproving other evidence). But even if it is all of these things, if the diligence of the party trying to reopen the record has to be weighed against the prejudice reopening the record will do to the other party. This final weighing test demonstrates the high importance we place on parties being able to question and examine evidence in a usual proceeding.
So, the IRS has the “golden ticket” (i.e. a document that shows actual supervisory approval) but the record is closed. Is that golden ticket enough to reopen or is the petitioner so prejudiced by this inability to confront the evidence that it should remain closed?
Clearly the supervisory approval form is very important to the case, meeting tests (1), (2) and (3) above. Further, the supervisory approval form is admissible as a hearsay exception through the business records rule (FRE 803(6)). However, the IRS supervisor declaration authenticating the supervisory approval form does, potentially, run afoul of the rules of evidence since it is offered after trial without reasonable written notice to the adverse party. See FRE 902(11).
In the end, the petitioners concern with being unable to challenge the supervisory approval forms is given little weight. Cross-exam would likely have done nothing. The issue is supervisory approval, which is shown by a particular form that the IRS is now offering: either the forms “answer those questions or they don’t.”
This seems like a practical way to frame an increasingly thorny issue.